Everyone wants to slow AI down, but the rules are still missing

The big story of the week is that some of the AI industry's most powerful voices suddenly agree on something: the pace needs to come down. Anthropic CEO Dario Amodei has called on the entire industry to slow down, and within a few hours OpenAI, Google DeepMind, Microsoft and Elon Musk backed the message. But not everyone is on board: Mark Zuckerberg refuses to put the whole industry on pause and believes each company is responsible for safe development itself, while Meta is simultaneously launching a new model.

The desire to slow down is not only a tech matter either. In Washington, left-wing Senator Bernie Sanders and right-wing Steve Bannon are joining forces in a call for tougher rules for AI. Political disagreement is thus temporarily set aside when it comes to artificial intelligence. But the global picture is not uniform: the Chinese state media outlet Global Times dismisses the slowdown proposal as a "Cold War playbook" aimed at China, a sign that the AI race is at least as much about geopolitics as about safety.

Behind the words, the rules lag

While top executives talk about caution, reality already shows why it is urgent. The Spanish data protection authority has received its first report of a data breach in which an AI agent (a program that can plan and carry out tasks on its own without constant human instruction) found a security flaw on its own, logged in and changed personal data. European Commission President Ursula von der Leyen calls such incidents only a foretaste of what is coming and wants to use the EU's AI Act to set global safety standards.

Here at home, legislation is also somewhat behind actual use. The government is trying again with a bill that would give authorities a clear right to use AI with citizens' personal data, but the Danish Data Ethics Council and a lawyer point out that clear rules for oversight of the use are still missing. It is the same pattern seen internationally: the rhetoric of responsibility is there, but concrete legislation lags behind.

What to watch for

This week's stories point in several directions at once: broad agreement that something must be done, but disagreement about who should do it and how quickly. Keep an eye on whether the big AI companies' talk of caution leads to concrete rules, or whether it mostly remains words. And watch whether the Danish bill on AI in the public sector ends up with clear oversight rules, or whether the uncertainty raised by the Data Ethics Council persists.